Real-Time Experiment Data

Contains Real-Time Experiment Data + Measurement Information of all Subjects.

Download the "All_Subjects_Results.zip" zip file, to get access to all data at once.

Structure:
Subject*_Results_Summary.txt (.txt file: Experiment Overview/ Settings + Individual & Global Model and Buffer Performances)
Subject*_Results.mat (.mat file: Recorded Experiment Data)

Real-Time Calibration Data

Contains Real-Time Calibration Data + Measurement Information of all Subjects.

Download the "All_Subjects_Calibration.zip" zip file, to get access to all data at once.

Structure:
Subject*_CalibrationSummary.txt (.txt file: Calibration Overview/ Settings)
Subject*_CalibrationLog.mat (.mat file: Recorded Calibration Data)

Real-Time Calibrated Models

Contains all subject-specific Calibrated Models of the Real-Time Experiment. (Base models calibrated/ retrained with the subject-specific Calibration Phase Data)

Download the "RT_All_Subjects_CalibratedModels.zip" zip file, to get access to all data at once.

3 individual models per subject:
Subject*_CalibratedModel_STEW/HEAT/MATB.mat
each model contains the corresponding CSP filters derived from the respective dataset (W_CSP_STEW, W_CSP_HEAT, W_CSP_MATB)

Real-Time Base Models

Contains the 3 Base Models chosen from the Offline Model and Calibration Analysis.

The 3 Base Models applied to the Real-Time Experiment are:

Base+CSP Standard STEW Model: "v1_Base_1000_25wCsp_4sec_proc5_STEW_model.mat"
Base+CSP Hyper HEAT Model: "v1_Base_hyper_1000_25wCsp_4sec_proc5_HEATCHAIR_model.mat"
Base+CSP Hyper MATB easy-meddiff Model: "v1_Base_hyper_1000_25wCsp_4sec_proc5_MATB_easy_meddiff_model.mat"

Each .mat file also contains the respective, dataset-specific CSP Filter "W_csp".

Offline Segmented, Labeled, Processed STEW Data

Raw, segmented STEW dataset (4-second raw STEW epochs)
Offline preprocessed STEW dataset (4-second proc5 STEW epochs)
STEW_labels (4-second epochs)

Offline Segmented, Labeled, Processed MATB Easy-MedDiff Data

Raw, segmented MATB Easy-MedDiff dataset (4-second raw MATB Easy-MedDiff epochs)
Offline preprocessed MATB Easy-MedDiff dataset (4-second proc5 MATB Easy-MedDiff epochs)
MATB Easy-MedDiff_labels (4-second epochs)

Offline Segmented, Labeled, Processed MATB Easy-Diff Data

Raw, segmented MATB Easy-Diff dataset (4-second raw MATB Easy-Diff epochs)
Offline preprocessed MATB Easy-Diff dataset (4-second proc5 MATB Easy-Diff epochs)
MATB Easy-Diff_labels (4-second epochs)

Offline Segmented, Labeled, Processed HEAT Data

Raw, segmented HEAT dataset (4-second raw HEAT epochs)
Offline preprocessed HEAT dataset (4-second proc5 HEAT epochs)
HEAT_labels (4-second epochs)

Offline Features & Models (Pre- & Post-Calibration)

Offline data of both the pre- and post-calibration offline pipeline.

Download the "Offline_Data.zip" zip file, to get access to all data at once.

"v1": first and only version

Note:
Labels: "4sec_sampled_labels.mat" - used for the Base Models and also later for the Calibration. (for feature set splitting into train, val, test) -> resulting in "4sec_train_labels.mat", "4sec_val_labels.mat", "4sec_test_labels.mat".

"AutoPipeline": 1000, 2000, 3000 and 4000 samples (4-sec epochs) as training data

• Base Models: ("training samples, feature configuration, 4-sec, preprocessing type, dataset name, model type")

"AutoCalibration": 1000, 2000, 3000 and 4000 training samples with respective 10, 25, 20, 25 and 30% Calibration Data (new target cross-data percentage of the training data on a base model)

• Offline Calibrated Base Models (Base Models calibrated with other Base Model Data (not subject data!). See: "_w*" --> Calibrated "w" other Base Model.